adding some fuzz to row height tolerance tests

Expanding on a4ab0c08e7 there's
another test that on some tinderboxes was a few permille off.

Adding 1% fuzz in comparison, seems we've now aquired some font
layout instabilities across platforms.

Change-Id: I7b068105c6ffad44b055e4dc37deb03c6040ab09
Reviewed-on: https://gerrit.libreoffice.org/55509
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>

sc/qa/unit/subsequent_filters-test.cxx

@@ -4067,9 +4067,9 @@ void ScFiltersTest::testTdf62268()
 
     SCTAB nTab = 0;
     nHeight = rDoc.GetRowHeight(0, nTab, false);
-    CPPUNIT_ASSERT_EQUAL(256, nHeight);
+    CPPUNIT_ASSERT_LESSEQUAL( 3, abs( 256 - nHeight ) );
     nHeight = rDoc.GetRowHeight(1, nTab, false);
-    CPPUNIT_ASSERT_EQUAL(1905, nHeight);
+    CPPUNIT_ASSERT_LESSEQUAL( 19, abs( 1905 - nHeight ) );
 
     xDocSh->DoClose();
 }